Lee Merkhofer Consulting Priority Systems

Technical Terms Used in Project Portfolio Management (Continued)






























An organization's financial resources (e.g., cash) available for investment in projects or other uses.

capital allocation

The portion of the capital budgeting process that deals with the allocation of the organization's available financial resources across business units, programs, and projects. The goal is to allocate resources in such a way as to maximize the value derived from the expenditure of those resources. Capital allocation is difficult in part because of the many alternative ways that specified budget can be allocated across different entities and because of the difficulty of determining how much benefit would be produced under each option. Optimizing the allocation of capital is a primary goal of project portfolio management.

capital budgeting

The process used by an organization to select and plan the expenditures that it will undertake over some defined, upcoming time period, for example, the next quarter, next year, or next five years. Such expenditures might include investments in property, plants, equipment, R&D, advertising, and so forth. For project based organizations, a key component of capital budgeting is deciding which specific projects to conduct and how much to spend on each. Since some expenditures may be quite large, capital budgeting may include determining how to finance those expenditures. These are important decisions, as they impact risk and the future success of the organization.

Traditionally, to support the selection of projects for the capital budget, organizations have mainly relied on techniques for estimating the incremental financial benefits available from each proposed project, expressed, for example, as a net present value (NPV), internal rate of return (IRR), profitability index (PI), and payback period. Project portfolio management (PPM) can improve on traditional project selection techniques by accounting for other types of project benefits in addition to financial benefits and by identifying project portfolios that collectively generate the most value for their costs. Organizations that are implementing PPM typically do so such that portfolio analyses are completed and the recommended project portfolio is identified for decision makers as the capital budget is being defined (see this paper for a description of how PPM is implemented in support of capital budgeting).

cardinal utility

A measure of an individual's subjective preference or satisfaction expressed numerically on a cardinal scale. This means that the utility numbers assigned (e.g., 1, 2, 3, etc.) are measured with respect to a benchmark scale, a scale such that both the numbers and differences between the numbers are meaningful measures.

A cardinal utility measurement is analogous to a measurement of temperature using a thermometer—both are expressed on cardinal scales. A thermometer measures temperature on a benchmark scale because the numbers for temperature obtained from the thermometer tell you both whether today is hotter than yesterday and by how much. It doesn't matter whether the thermometer measures in Fahrenheit or centigrade, because the two are linearly related (degrees Fahrenheit equals degrees centigrade times 1.8 plus 32). Likewise, a cardinal utility measure of preference will tell you both whether you prefer one thing to another and by how much. Like the number obtained using a thermometer, the utility number assigned is unique up to a linear transformation. Contrast cardinal utility with ordinal utility.

cardinal utility function

A utility function that expresses utility on a cardinal scale.

cardinal value function

A value function that expresses utility on a cardinal scale. A cardinal value function is also called a measurable value function.

As further explanation, a value function is a special case of a utility function applicable only to conditions of certainty (i.e., when there is no uncertainty regarding the levels of the utility function's attributes). Thus, a cardinal value function is the special case of a cardinal utility function applicable when there is no uncertainty over the levels of the attributes.

certain equivalent

Also called certainty equivalent or risk-adjusted value, a measure of the value of an alternative or outcome that takes uncertainty (risk) into account. The certain equivalent of an alternative is the amount that the decision maker would be indifferent between (1) having that monetary amount for certain or (2) having the alternative with its uncertain outcome. The certain equivalent for a project is typically less than its expected value and depends in part on the decision maker's willingness to accept risk. For example, a risk averse decision maker might have a certain equivalent of $500,000 for a project with equal chances of yielding $0 and $2,000,000, even though the expected value for this project would be $1,000,000.

Decision theory, which provides a means for encoding a decision maker's preferences in a mathematical utility function, provides a means for calculating the certain equivalent. A utility function U provides a number that quantifies how satisfied the decision maker would be, depending on the outcome. These utility functions can be indexed by risk tolerance, which quantifies the decision maker's willingness to accept risk. The greater the decision maker's risk tolerance, the closer the certain equivalent of a gamble will be to its expected value.

Among the useful properties of the certain equivalent is the fact that the certain equivalent of the sum of lotteries is the sum of the certain equivalents of each lottery, provided that the uncertainties are independent (not correlated). This means that, just as the expected values of uncertain independent projects may be added to obtain the expected value of the project portfolio, if the decision maker is risk averse, the certain equivalent of each project may be added to obtain the certain equivalent of the project portfolio.

characteristic objects method (COMET)

A multi-criteria analysis (MCA) decision-making method based on paired comparisons and concepts of fuzzy logic. A criticism of many MCA methods using pairwise comparisons is rank reversal—adding a new alternative to the original set can cause the ranking of the other alternatives to change. COMET avoids this problem by applying pairwise comparison not to the original set of alternatives but to a larger set of constructed alternatives called characteristic objects which are independent of the actual decision alternatives. Pairwise comparisons with respect to each of the criteria are used to obtain a relative preference measure for each of the characteristic objects. Preferences for each of the original decision alternatives are then obtained based on a mathematical concept of its distance from the nearest characteristic objects and their preference values.

clairvoyant test

A test that may be used to verify that a performance measure used to compare alternatives is an observable; that is, capable of being observed and measured. Being observable is a desirable characteristic for the measures used in a decision model.

Suppose you are a decision maker in a company considering a change to one of your products. One objective might be satisfying your customers. Is "customer satisfaction" a useful performance measure?

To conduct the test, imagine the existence of a clairvoyant (someone who can foresee the future). Assume you decide to make the specified change to you product. Ask the clairvoyant to tell you the level of customer satisfaction that will result. Could the clairvoyant answer your question? The answer is no, because the clairvoyant would need to ask you to define customer satisfaction and how you want to measure it. If, on the other hand, your performance measure was "the number of customer complaints received" or your company's ranking in the next customer satisfaction survey, the clairvoyant could tell you that. Performance measures that fail the clairvoyant test are not observables. Performance measures that aren't observables are ambiguous or vague so should not be used in decision models.


As in client-server, see Server.


A communications network. The word most often refers to the Internet, and more precisely to some data center with servers that is connected to the Internet.

cloud computing

An overused buzzword that can mean different things. Originally, the term referred to a technology that mimics supercomputing capability (high-speed mathematical calculations) by simultaneously applying the computing resources available from numerous servers available on a network. The concept is to divide a computational task into pieces and use specialized connections to allow groups of servers (typically those involving low-cost PC technology) to simultaneously conduct the data processing chores.

Today, the term most often refers to a method for delivering information technology services over the internet through web-based tools, rather than through personal computers or direct connections to local severs. Cloud computing relies on sharing computing resources. Rather than having to build and maintain computing infrastructure in-house, organizations can purchase the desired computational resources as a utility, just as they purchase water and electricity.

Some project portfolio management tool providers, and other software vendors, are using the term in a more mundane way as a synonym for software as a service (SaaS)—their applications are available via the Internet cloud provided from their Web servers and accessed on demand via the user's Web browser.

cognitive dissonance

Anxiety one feels from simultaneously holding contradictory or otherwise incompatible thoughts or beliefs, as when one likes a person but disapproves strongly of one of his or her habits

comfort zone biases

A category of related cognitive biases with the common characteristic that their effect is to promote behavior that is comfortable rather than reasoned. Examples of comfort zone bias include status quo bias, supporting evidence bias, and sunk cost bias.


The process and steps for introducing a new product into the marketplace. Decisions around commercialization are often critical to market success, including when, where, and how to launch the product. Commercialization is often the most expensive component associated with projects designed to create new consumer products, as it typically includes the cost of advertising, sales promotion, and other costly marketing efforts.

compensatory method

A multicriteria analysis, decision-making method with the characteristic that poor performance relative to one or more criteria can be compensated for by good performance relative to other criteria. Compensatory methods typically involve four steps:

  1. Specifying criteria for evaluating alternatives
  2. Assigning weights to indicate the importance of the criteria
  3. Judging the performance of each alternative with respect to each criterion
  4. Applying an aggregation equation or mathematical algorithm to combine the weights and performance judgments to compute a "desirability" measure (number) for ranking the alternatives

Examples of compensatory methods include multi-attribute utility analysis (MUA), the analytic hierarchy process (AHP), the evidential reasoning approach (ER), and most scoring models.

Project portfolio management tools with the capability to prioritize project almost always use a compensatory method.

compromising method

Also called compromise method, a sub-category of multi-criteria analysis (MCA) decision-making methods composed of methods that rank the alternatives to a decision based on calculating how "close" the alternative is to an ideal solution and how "far" it is from the nadir solution. The ideal alternative is defined as a hypothetical alternative that performs at the best possible level on every criterion. The nadir alternative is defined as a hypothetical alternative that performs at the worst possible level on every criterion. "Distance" is defined using the natural extension to the concept of distance in 3-dimensional space to the multi-dimensional space defined by the number of criteria. VIKOR and the Technique for Order Preference by Similarity to Ideal Solution (TOPSIS) are examples of compromising methods.

Methods in this category are termed "compromising" based on their interpretation when applied to high-conflict, group decision making situations. Presumably, stakeholders can agree on the criteria for evaluating alternatives. Thus, everyone should agree on the ideal and nadir solutions. An alternative that is as close as possible to the ideal and as far as possible from the nadir should be one that achieves a high level of agreement. The group as a whole should be pleased that the selected solution is close to the ideal. Those opposing the choice should presumably feel minimal regret based the choice being so far from the worst possible choice.

concordance method

Also called concordance analysis a sub-category of multicriteria analysis (MCA) methods based on identifying and comparing criteria that support a particular alternative (concordance) alongside the criteria that oppose the alternative (discordance). (Concordance means agreement, concord, harmony.) Concordance methods are a type of outranking method. The various ELECTRE methods and PROMETHEE are the most used and well-known of the concordance methods.

Like other outranking methods, with concordance methods each pair of alternatives is compared to determine whether one alternative outranks another (performs as well or better on every criterion). Once all alternatives have been compared, the results are combined to obtain a partial (or complete) ranking order.

Concordance methods are non-compensatory in that there is no need to assess strengths of preference or tradeoffs among the criteria; all that is required are the ordinal results from the pairwise comparisons. The limitation is that the resulting ranking is ordinal rather than cardinal, meaning that there is no ability to divide a utility number by cost to obtain a ranking that maximizes portfolio value..

confidence interval

Also called subjective confidence interval or prediction interval, a range of possible values defined for an uncertain quantify such that there is some specified probability (e.g., 90%) that the true or actual value of the quantity lies within the range. In an confidence interval, the width of the interval describes how uncertain we are about an unknown quantity.

Data correlation

An example confidence interval

The term has a somewhat different meaning in statistics, where it describes an interval or range for some statistic computed from a sample of observations from some population (e.g., the mean value) such that the "true" population statistic can be expected to be located within that interval with specified level of certainty (e.g., 90%).

conjoint analysis

A method often used in the field of market research to quantify how people value the various features or attributes that make up a product or service. Subjects are asked to state their preferences for choices presented as paired comparisons of product choices defined by two or more attributes, for example, a low cost computer with 2 GHz processor, 512 MB RAM, and a 15 inch monitor versus a higher cost computer with 4 GHz processor, 2 GB RAM and a 21 inch monitor. In some versions, the subjects indicate their strength of preference, for example, on a 1 to 10 scale. The approach requires answers to be provided for many such questions, however, since most people find the questions easy, conjoint studies can often be conducted by mail survey or over the internet.

The data obtained from a conjoint analysis may be used to derive a utility function for quantifying customer preferences for products with various combinations of attributes. A specific mathematical form for the utility function is assumed (e.g., additive) and the data from the conjoint analysis is used to set the parameters of the function so as to best fit the data. The utility function may then be used to predict future customer choices. The technique may be used to quantify preferences for items other than products, and some project prioritization tools use conjoint analysis as a method for ranking projects.

connotative scale

A type of a scoring scale designed to measure the connotative meaning of objects, events, or concepts. The connotations are used to derive a subject's attitude towards the given object, event or concept scale. On this website, the term typically refers to a decision to conduct a project. Some projects, for example, may have positive, negative, or neutral connotations.


An outcome of a decision. On this website, the term typically refers to a decision to conduct a project. Consequences are outcomes that matter to the decision maker; that is, an outcome that the decision maker cares about. Consequences are typically quantified via performance measures that indicate the degree to which the decision maker's objectives are achieved.

consequence model

A model for simulating or otherwise relating results or outcomes that people care about to specified actions and conditions. Project portfolio management tools that prioritize projects based on the likely consequences of those projects contain consequence models. Consequence modeling is the process of constructing a consequence model and typically involves developing a description of cause-effect relationships or linkages. Various techniques are available for creating consequence models, and many such models exist and have become well-established within various disciplines and application areas.


In risk analysis, a policy of assuming pessimistic outcomes for uncertainties when quantifying risk. Conservatism reflects a preference for erring on the side of overstating as opposed to understating risk in situations where uncertainties complicate the quantification of risk. For example, if a probability distribution for an uncertain undesired outcome is known, conservatism might involve selecting an outcome estimate at, say, the 95th percentile—meaning there is a 95% chance that the actual outcome will be overestimated (the outcome will be more desirable than assumed) and only a 5% chance that it is underestimated (less desirable than assumed). By substituting conservative estimates for probability distributions, conservatism allows for a simpler, deterministic analysis of risk. Conservatism is criticized, however, for injecting a bias into risk analysis. Also, studies show that conservative assumptions tend to compound with the result that estimated risks are far larger than an objective analysis would report.


In the mathematical sense, a constraint is a condition of an decision problem that the solution must satisfy. There are several types of constraints—primarily equality constraints, inequality constraints, and integer constraints. The set of candidate solutions that satisfy all constraints is called the feasible set Constraints may be dictated by the environment, physical processes, economics, cultural, legal and/or resources aspects which must be satisfied in order to obtain an acceptable solution. As one example, conducting all projects required by law is a common constraint imposed on the choice of a project portfolio.

constrained optimization

A class of mathematical solution techniques aimed at optimizing an objective function with respect to variables some of which are subject to constraints. The techniques for constrained optimization include linear programming, non-linear programming, integer programming, dynamic programming, and multi-objective programming.

constructed metric

Also called a constructed measure, a measure constructed in terms of other measures. Examples include the Dow Jones Industrial Average, the Consumer Price Index, and the Michelin 3-star measure for rating restaurants. Constructed measures are typically defined when neither natural measures nor proxy measures cannot be found. Sometimes, the term is used to refer to a measure based on a constructed scale.

constructed scale

See attribute and scale.

consumer surplus

The difference between the total amount that consumers would be willing to pay and the total amount that they actually do pay (i.e. the market price). In other words, consumer surplus is the value consumers receive over and above what they actually have to pay for goods and services. Consumer surplus is a type of value obtained by customers when organizations provide products or services that are purchased by customers.

contingency fund

Resources set aside for expenses associated with unexpected but potential project or other business setbacks. For example, a contingency fund might be established in the context of project portfolio management (PPM) to provide additional funding for projects that might go over budget (although the PPM process would include reevaluating such projects to see if the additional funds are warranted and what changes should be made to the project plan in light of a budget overrun).

contingent valuation

A method, generally associated with cost benefit analysis, for assigning monetary values to consequences based on asking individuals about their willingness to pay to obtain desired outcomes or to reduce or willingly accept adverse outcomes. Contingent valuation is often used to value environmental consequences such as increased levels of pollution or noise.

continuous risk, continuous uncertainty

A risk or uncertainty with that will produce an outcome within a range of possibilities. Continuous uncertainties are characterized by continuous probability distributions (e.g., probability density functions) that indicate the relative probability of the possibilities. For comparison, see discrete risk.

continuous scale

A measurement scale that allows assigning an infinite number of possible number values, for example, any number between zero and one.


A statistical relationship between two or more variables such that systematic changes in the value of one variable tend to be accompanied by systematic changes in the other. For example, as illustrated below, the height of parents and mature children can be shown to be correlated based on plotting on a graph parent heights versus child heights. If there was no correlation, the plot would appear as dots spread across a roughly circular shape. Instead, the plot appears as an oval shape aligned along approximately a 45 degree line from the origin of the graph. The shape demonstrates that taller parents tend to have taller children, but the relationship is not perfect, indicating that there are other factors involved.

Data correlation

A sample correlation plot (parent vs. children height)

Correlation may be quantified by calculating a coefficient of correlation, denoted ρ, which is a number between -1 and 1 that indicates the strength and direction of a linear relationship between two variables. Regression analysis may be used to compute correlation coefficients.

cost benefit analysis

Also called benefit cost analysis, a decision making theory and collection of related analytic techniques for evaluating decisions based on comparing benefits and costs. While many decision-aiding tools claim to do this, CBA is distinguished by its foundation in theory and by its rigorous process for computing costs and benefits. Like decision analysis (DA), CBA is essentially a "megatool;" a coherent set of concepts and techniques that may be used to identify, estimate, and place monetary values on the impacts of proposed actions.

A distinct characteristic of CBA, compared to many other decision aiding approaches, is that it does not quantify the costs and benefits of actions from the specific perspective of responsible decision makers. Instead, CBA measures costs and benefits from the perspective of society at large. Thus, CBA is most applicable for government decision making. CBA requires identifying all parties affected by a proposed action and estimating the monetary value of the effects the action would have on their welfare. The US government began using CBA in the 1930s, and CBA continues to be the mostly widely used approach for formally evaluating and prioritizing major projects undertaken by government agencies.

With CBA, costs and benefits are measured relative to a "do-nothing," status quo option. According to CBA, an action should be considered only if its net benefit (benefit minus cost) is greater than zero. According to CBA, all alternatives with positive net benefit should be considered, and the best alternative is the one that leads to the greatest benefit gain (i.e., the alternative with the largest net benefit).

To determine the monetary value of the impacts of proposed projects, CBA relies on the concept of market prices. In theory, a free market generates prices by balancing aggregate demand with aggregate supply. Each individual adjusts his or her purchases until the value of the last item purchased is just worth what it costs. Thus, the prices that result indicate the marginal benefit realized from each individual's consumption of each good.

Following this logic, CBA attempts to use market prices to value project impacts. For example, suppose a project is proposed to clean up a hazardous waste site. CBA might use real estate prices to estimate the value of the cleanup effort. The market values of similar properties close to and far from the waste site would be compared to determine the value loss suffered by those living near to the site. Removing the site would, presumably, eliminate the property value differences and create this much added value for home owners. This monetary value could be compared to the costs of the project to determine whether the project should be conducted. To value project impacts for which no market exists, CBA uses procedures that indirectly reference market prices. For example, values for CBA are often obtained based on surveys or interviews with people to estimate their willingness to pay to obtain or avoid the effects in question.

Although CBA is widely used and based on compelling theory for decision making, the approach is considered by many to be controversial. CBA ignores the way in which the costs and benefits of proposed actions are distributed, a consideration that is often important to decision makers. Also, numerous applications of CBA have demonstrated that the approach frequently concludes that the benefits of proposed government interventions do not justify the costs. This result may be due the fact that some project benefits simply cannot be addressed through reference to prices that exist in the marketplace.

Since CBA does not normally rely on methods for assessing and incorporating expert judgments, it can fail to account for project outcomes that do not have immediate, tangible, economic implications. Likewise, CBA may fail to address important risks in situations lacking data for quantifying uncertainties. CBA provides little opportunity for stakeholders to contribute to the evaluation process, except perhaps, in framing the problem (e.g., identifying alternatives). On the other hand, CBA avoids the necessity of decision makers providing subjective value judgments. It therefore appeals to some because it appears to be a more value-free guide to decision making. In truth, though, CBA embodies strong value judgments.

cost effectiveness analysis (CEA)

A form of economic analysis that compares the costs of alternatives for achieving the same or similar outcomes. CEA can be viewed as a version of cost-benefit analysis (CBA) that does not include monetizing the estimated benefits of alternatives, and, as such, it is often used in situations where it would be difficult or uncomfortable to put a dollar value on outcomes.

For example, "number of lives saved" might be an outcome (or performance measure) for investments in health care. In order to compare or prioritize health care investments, CBA includes techniques for expressing outcomes in equivalent monetary units so that the values can be combined and compared with costs. (Other decision models also express the value of various outcomes in common units, though the unit need not be dollars.) CEA might be preferred over CBA if assigning an explicit monetary value to saving a life is viewed as too controversial.

Though CEA may be less controversial than CBA, the technique is not so helpful if the alternatives under consideration have different costs and would produce different outcomes. If, for example, more lives would be saved under a more expensive alternative, then decision makers must judge whether the additional lives saved from the more expensive alternative would be worth the additional costs. In the absence of a model that allows assigning explicit values, those tradeoffs are made implicitly, so the logic for choice is less transparent. For this reason, CEA is most useful when the alternatives under consideration have different costs, but would produce the same outcomes.

cost of capital

The cost a company effectively pays in order to obtain cash (from debt and equity sources) to finance its operations. It is the return that is required on company investments to compensate the investor, taking into account the risk involved. The cost of capital is generally calculated on a weighted average basis (see WACC).


Any quality, property, characteristic, or rule established to guide decision making. The term has somewhat different meaning in different contexts, as attributes, objectives and goals may all be referred to as criteria. Reflecting this, the term a multicriteria decision problem means either a multi-attribute or a multiobjective decision problem (or both). The term multicriterion decision making (MCDM) is, therefore, used to indicate the general field of study which includes decision making in the presence of two or more conflicting objectives and/or decision analysis processes involving two or more attributes. The use of multiple criteria for decision making is the subject of multi-criteria analysis.

critical path

The subset of activities within a project whose duration determines the duration of the project. If an activity along the critical path is delayed, then the project will be delayed.

cumulative probability distribution

A probability distribution expressing the probability that a random variable X will take a value less than or equal to x. The figures below provide examples of cumulative probability distributions. The first is as sample cumulative probability distribution for a discrete uncertainty , the second is for a continuous uncertainty.

Cumulative probability distributions

Cumulative probability distributions (outcome from a pair of dice and normal distribution)


The internal rate of return (IRR) for a project below which the project is considered unacceptable. It is often taken to be the project opportunity cost or the organization's cost of capital. The cut-off rate would be the minimum acceptable irr for a project or the discount rate used to calculate the project's net present value.